AI Model Shows Promise for Mapping Periodontal Bone Defects on Radiographs

Start Discussion
Posted: June 23, 2026

AI Model Shows Promise for Mapping Periodontal Bone Defects on Radiographs

Edited by Dentaltown staff

A deep learning model identified, classified, and outlined infraosseous periodontal defects on dental radiographs with good overall accuracy, though its developers said larger datasets are needed before clinical use, according to a study published June 23 in BMC Oral Health.

Researchers trained a YOLOv8 neural network on 329 periapical radiographs, each manually annotated to mark the bone defects around teeth. Such intrabony defects are difficult to assess on two-dimensional images because overlapping anatomy can obscure their shape.

The model reached an area under the ROC curve of 0.81, a measure of its ability to distinguish defect types, which the authors characterized as good discriminative power. Its classification and segmentation performance was more modest.

Agreement between the model’s outlines and the human annotations was strongest for one-wall, three-wall, and four-wall defects, with Dice similarity scores near 0.87 for each.

Defect morphology influences a tooth’s prognosis, surgical planning, and regenerative potential, the authors noted, making accurate classification clinically relevant for periodontal and surgical care.

The authors said the model’s performance was limited by dataset size and that future work would focus on building larger, more refined training sets.

The study was led by Daniel Palkovics and colleagues in the Department of Periodontology at Semmelweis University in Budapest, Hungary. The work adds to a growing body of research on neural networks for interpreting dental imaging.

Sources:
BMC Oral Health, “Deep learning-based identification, classification and segmentation of infraosseous periodontal defects using a YOLOv8 neural network,” June 23, 2026, DOI 10.1186/s12903-026-08919-x: bmcoralhealth.biomedcentral.com


AI Model Shows Promise for Mapping Periodontal Bone Defects on Radiographs

Views: 6
Start Discussion
Dentaltown Magazine
June 2026
Dentaltown Magazine June 2026 Read Current Issue Read Archived Issues
Sponsors
Townie Perks
Townie® Poll
What part of a dental office do you feel makes the strongest first impression on patients?
  
Site Menu
Message Boards Magazine Continuing Education Podcasts Blogs Classifieds Dental News Events Townie® Poll Ebooks Townie Choice Awards
Company Info
Contact Us Advertise With Us Howard Farran DDS, MBA Policies and Agreements
Site Help
The Dentaltown Team, Farran Media Support
Phone: +1-480-445-9710
Email: support@dentaltown.com
©2026 Dentaltown, a division of Farran Media • All Rights Reserved
9633 S. 48th Street Suite 200 • Phoenix, AZ 85044 • Phone:+1-480-598-0001 • Fax:+1-480-598-3450